Premium
Stress Testing of Financial Industries: A Simple New Approach to Joint Stress Testing of Korean Banking, Securities, and Non‐Life Insurance Industries *
Author(s) -
Chang KookHyun,
Kim MyungJig
Publication year - 2009
Publication title -
asia‐pacific journal of financial studies
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 15
eISSN - 2041-6156
pISSN - 1226-1165
DOI - 10.1111/j.2041-6156.2009.tb00022.x
Subject(s) - stress test , business , stress testing (software) , joint (building) , financial ratio , stress (linguistics) , actuarial science , finance , computer science , engineering , architectural engineering , linguistics , philosophy , programming language
This paper proposes a simple joint stress testing model useful in studying the effects of specific stress scenarios on a financial sector. In doing so, we adopt the principal component analysis (PCA) as a main device to interpret various financial information contained in figures and numbers on a financial company. We repeat the principal component analysis across different levels from individual company to a financial industry, and eventually to a financial sector as a whole to derive a financial sector risk index. We then link the sector risk index with stress macro variables, which constitute a much simpler task than devising individual models for each financial components. Once a relationship is established, a joint stress test is conducted by repeating PCA conversely. As a sample of stress scenario in the paper, we use the case of the 2003 credit card distress. We find that securities industry is more sensitive to market stresses than two other industries‐bank and insurance‐and that financial institutions in such a stress‐sensitive industry are, consequently, more affected by the stresses than those in other industries. Despite the simplicity of the proposed model, this model is expected to provide substantial information, particularly for financial supervisors without having to build a complicated joint stress testing model.